Arabic NLP researcher behind TTLab shared-task systems for stance detection and MT error-span detection
2
papers here
0
reviewed by Evolve
1
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-09-27) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
No output score is available: the bibliometric snapshot does not yet provide enough indexed publications to measure productivity. This axis would count total and annual publications and their venues once a fuller record is linked.
Impact—
Insufficient citation data prevents scoring impact. This axis would measure citations, citing works, and citation-weighted influence of the publications; the two reviewed preprints are too recent for meaningful citation accumulation.
Focus—
The topical focus score is not computable from current data. It would measure concentration of output across fields and keywords; the two reviewed papers both sit in Arabic NLP, but the indexed record is too thin to generalize.
Collaboration28/100
Scores 28/100, computed from a raw count of 3 distinct coauthors in the bibliometric record. The two reviewed papers are TTLab team entries for shared tasks, with Alexander Mehler and Bhuvanesh Verma recurring among frequent coauthors; Ali Abusaleh is first author on one and a middle author on the other, neither corresponding. The pattern suggests a small, stable working group rather than a wide collaboration network.
Activity—
No activity score is available yet. This axis tracks recent publication cadence and year-over-year output; with only 2026 shared-task papers in scope, there is no multi-year signal to evaluate.
Reproducibility—
Reproducibility is unscored for now. It would assess code, data, and artifact availability of the papers; Evolve's review flags both systems at medium repro-readiness, but this qualitative note is not a computed bibliometric score.
Research summary
Ali Abusaleh is an Arabic natural language processing researcher whose Evolve-indexed record centers on competitive shared-task system papers. In 2026 they co-authored two TTLab submissions released as arXiv preprints. The first, at StanceEval-2026, applies a cloze-style prompting approach (CLASP-Ar) for Arabic-language stance detection, using verbalizer-constrained masked language modeling and pattern-exploiting training; it reached 71.36/74.14 Favg2, with ablations the review calls useful, though encoder conflicts and a dev-test gap remain unexplained. The second, first-authored at AlexandriaX-2026, is a fine-tuned surface tagger for Arabic machine-translation error-span detection and classification, trained with focal loss to handle class imbalance; it ranked 3rd with 40.91/42.00, with honest ablations but near-zero F1 on rare error classes. Both papers are team efforts within a small collaboration footprint: the bibliometric record lists 3 distinct coauthors, with Alexander Mehler and Bhuvanesh Verma recurring. Neither paper lists Abusaleh as corresponding author. Key themes include Arabic and dialectal language technology, prompting-based and tagging-based methods, and evaluation under class imbalance. Substantive caveats apply: the profile is name-resolved from public bibliographic data without institutional or ORCID confirmation, and five of six radar dimensions (output, impact, focus, activity, reproducibility) currently lack sufficient data, so this should be read as an early-stage profile that will sharpen as the indexed record grows.
Collaboration pattern
The visible collaboration footprint is compact: 3 distinct coauthors in the bibliometric record, with Alexander Mehler and Bhuvanesh Verma recurring across the reviewed work. Both papers are TTLab team submissions to shared tasks, and Abusaleh moves between first author (AlexandriaX-2026 tagger) and middle author (StanceEval-2026 stance system), never corresponding author — a pattern consistent with contributing inside a small, stable lab group rather than leading an independent network.
Research keywords
Highlights
Cloze-style prompting for Arabic stance detection at StanceEval-2026
First-authored Arabic MT error-span tagger ranks 3rd at AlexandriaX-2026
A shared-task-centered, ablation-driven research pattern
Both 2026 papers are TTLab team entries to Arabic NLP evaluation campaigns, pairing competitive scores with ablation studies; Evolve's reviews consistently flag unexplained gaps — a dev-test discrepancy on one and near-zero rare-class F1 on the other.
Timeline
- 2026Co-authored TTLab's StanceEval-2026 system paper on Arabic stance detection via cloze-style prompting (CLASP-Ar), in a middle author position.
- 2026First-authored TTLab's AlexandriaX-2026 paper on Arabic machine-translation error-span detection and classification.
- 2026Recorded a 3rd-place result (40.91/42.00) in the AlexandriaX-2026 error-span shared task with a focal-loss surface tagger.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Radar scores are computed from a limited OpenAlex-based bibliometric snapshot; five of six dimensions (output, impact, focus, activity, reproducibility) currently lack sufficient data.
- This profile is name-resolved from public bibliographic data; no institution, country, or ORCID was available to confirm the researcher's identity.
- Coverage is limited to the two arXiv preprints reviewed by Evolve; preprints are not peer-reviewed, and other work by this researcher may fall outside this scope.
- The collaboration score of 28 reflects only 3 distinct coauthors in the available record and may understate collaboration captured by fuller indexing.